Last month I opened an AI invoice and had that tiny stomach-drop moment every solo dev knows.
I was using a handful of models across a bunch of tabs and tools, and the cost was not obvious until the bill showed up. Nothing was broken. It was just invisible.
That bugged me enough that I started building TokenBar.
TokenBar lives in the menu bar and shows token usage in real time so I can see what I am actually spending while I work. Not after the fact. Not buried in a dashboard I never check. Right there, while the context is still fresh.
The lesson was not really about tokens. It was about visibility.
When a cost is hidden, I ignore it.
When a cost is visible, I make better decisions.
I think the same thing applies to attention, too. If my phone is buzzing, Slack is open, and a dozen browser tabs are stealing context, I do not feel distracted in some abstract way. I just feel slower. The damage shows up later in the work.
That is the other reason I like building small tools. They turn vague pain into something concrete enough to fix.
TokenBar is for the bill.
Monk Mode is for the distraction.
Both came from the same place: I wanted fewer surprises and more control over the thing actually costing me money or time.
If you are building with AI and you have not looked closely at your usage lately, do it now. You might not need another model. You might just need a better dashboard for the one you already have.
That was my wake up call anyway.
Top comments (0)